perf: store splat arg index in a single byte by teor2345 · Pull Request #158251 · rust-lang/rust

teor2345 · 2026-06-22T14:20:02Z

This is a perf test for a variation of #153697.
It's part of the splat lang experiment, tracking issue #153629.

And remove redundant const fn

- const, async, and unsafe functions - const generics, complex types, and where clauses (including impl Tuple)

JonathanBrouwer · 2026-06-22T14:22:52Z

perf: store splat arg index in a single byte

rust-bors · 2026-06-22T16:37:06Z

☀️ Try build successful (CI)
Build commit: 51cfa71 (51cfa71c5e3563738f2ed8b82f18684e522dbc8d)
Base parent: 9030e34 (9030e345fe92df1ccefd0a8cdf61a9a9a5b73cb0)

rust-timer · 2026-06-22T17:24:14Z

Finished benchmarking commit (51cfa71): comparison URL.

Overall result: ❌ regressions - please read:

Benchmarking means the PR may be perf-sensitive. It's automatically marked not fit for rolling up. Overriding is possible but disadvised: it risks changing compiler perf.

Next, please: If you can, justify the regressions found in this try perf run in writing along with @rustbot label: +perf-regression-triaged. If not, fix the regressions and do another perf run. Neutral or positive results will clear the label automatically.

@bors rollup=never
@rustbot label: -S-waiting-on-perf +perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	0.1%	[0.1%, 0.1%]	5
Regressions ❌ (secondary)	0.2%	[0.1%, 0.3%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.1%	[0.1%, 0.1%]	5

Max RSS (memory usage)

Results (secondary -2.4%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.4%	[-3.3%, -1.7%]	6
All ❌✅ (primary)	-	-	0

Cycles

Results (secondary 4.3%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	4.9%	[1.7%, 18.5%]	11
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.0%	[-3.0%, -3.0%]	1
All ❌✅ (primary)	-	-	0

Binary size

Results (primary 0.0%, secondary 0.1%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	0.0%	[0.0%, 0.1%]	56
Regressions ❌ (secondary)	0.1%	[0.0%, 0.3%]	75
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.0%	[0.0%, 0.1%]	56

Bootstrap: 503.321s -> 509.023s (1.13%)
Artifact size: 353.40 MiB -> 353.69 MiB (0.08%)

teor2345 · 2026-06-22T17:48:44Z

0.1% regression on 5 primary and 0.2% on 4 secondary benchmarks seems like as good as we'll get.

I'll replace #153697 with this branch once I've checked with Oli.

teor2345 · 2026-06-23T06:27:10Z

I'm going to push these perf fixes to #153697

…r=oli-obk Add arg splat experiment initial tuple impl ### Description This PR is part of the argument splatting lang experiment, and FFI overloading / C++ interop project goals: - rust-lang#153629 - https://rust-lang.github.io/rust-project-goals/2026/overloading-for-ffi.html - https://rust-lang.github.io/rust-project-goals/2025h2/interop-problem-map.html Example code using existing unstable features: - https://play.rust-lang.org/?version=nightly&mode=debug&edition=2024&gist=f42a3754a63a3d9365670e57257053d5 Discussion of implementation strategy: - [#t-lang > On overloading @ 💬](https://rust-lang.zulipchat.com/#narrow/channel/213817-t-lang/topic/On.20overloading/near/579590336) The PR is the initial implementation of the feature: - `splat` incomplete feature gate - `#[splat]` attribute on function arguments - Splatted function argument TypeInfo - `#[splat]` function parameter check at THIR level - splatted MIR lowering (as tupled arguments) - feature gate and UI tests for item type filtering, non-splattable arguments, splattable tuples, generics, and the "overloading at home" example - about half the diff (1100 lines) is tests and test output Once this PR merges, we can add further functionality, then test it out in interop tools. ### Perf Impact We expect a 0.1% regression on 5 primary and 0.2% regression on 4 secondary benchmarks in this PR, based on [this perf run](rust-lang#158251 (comment)). We tried a number of different ways to improve perf. Limiting splat to the 255th or lower argument is a simple hack that gives good perf, and is good enough for an experiment. This PR series already has significant perf wins in rust-lang#155223 - [0.3% perf improvement across 45 primary benchmarks](rust-lang#155223 (comment)). We're spending a small amount of that perf for the new feature in this PR. ### Out of Scope for this PR - Change codegen to de-tuple caller and callee - Better diagnostics - Full support for splatted function pointer arguments

Rollup merge of #153697 - teor2345:fn-arg-splat-experiment, r=oli-obk Add arg splat experiment initial tuple impl ### Description This PR is part of the argument splatting lang experiment, and FFI overloading / C++ interop project goals: - #153629 - https://rust-lang.github.io/rust-project-goals/2026/overloading-for-ffi.html - https://rust-lang.github.io/rust-project-goals/2025h2/interop-problem-map.html Example code using existing unstable features: - https://play.rust-lang.org/?version=nightly&mode=debug&edition=2024&gist=f42a3754a63a3d9365670e57257053d5 Discussion of implementation strategy: - [#t-lang > On overloading @ 💬](https://rust-lang.zulipchat.com/#narrow/channel/213817-t-lang/topic/On.20overloading/near/579590336) The PR is the initial implementation of the feature: - `splat` incomplete feature gate - `#[splat]` attribute on function arguments - Splatted function argument TypeInfo - `#[splat]` function parameter check at THIR level - splatted MIR lowering (as tupled arguments) - feature gate and UI tests for item type filtering, non-splattable arguments, splattable tuples, generics, and the "overloading at home" example - about half the diff (1100 lines) is tests and test output Once this PR merges, we can add further functionality, then test it out in interop tools. ### Perf Impact We expect a 0.1% regression on 5 primary and 0.2% regression on 4 secondary benchmarks in this PR, based on [this perf run](#158251 (comment)). We tried a number of different ways to improve perf. Limiting splat to the 255th or lower argument is a simple hack that gives good perf, and is good enough for an experiment. This PR series already has significant perf wins in #155223 - [0.3% perf improvement across 45 primary benchmarks](#155223 (comment)). We're spending a small amount of that perf for the new feature in this PR. ### Out of Scope for this PR - Change codegen to de-tuple caller and callee - Better diagnostics - Full support for splatted function pointer arguments

Add arg splat experiment initial tuple impl ### Description This PR is part of the argument splatting lang experiment, and FFI overloading / C++ interop project goals: - rust-lang/rust#153629 - https://rust-lang.github.io/rust-project-goals/2026/overloading-for-ffi.html - https://rust-lang.github.io/rust-project-goals/2025h2/interop-problem-map.html Example code using existing unstable features: - https://play.rust-lang.org/?version=nightly&mode=debug&edition=2024&gist=f42a3754a63a3d9365670e57257053d5 Discussion of implementation strategy: - [#t-lang > On overloading @ 💬](https://rust-lang.zulipchat.com/#narrow/channel/213817-t-lang/topic/On.20overloading/near/579590336) The PR is the initial implementation of the feature: - `splat` incomplete feature gate - `#[splat]` attribute on function arguments - Splatted function argument TypeInfo - `#[splat]` function parameter check at THIR level - splatted MIR lowering (as tupled arguments) - feature gate and UI tests for item type filtering, non-splattable arguments, splattable tuples, generics, and the "overloading at home" example - about half the diff (1100 lines) is tests and test output Once this PR merges, we can add further functionality, then test it out in interop tools. ### Perf Impact We expect a 0.1% regression on 5 primary and 0.2% regression on 4 secondary benchmarks in this PR, based on [this perf run](rust-lang/rust#158251 (comment)). We tried a number of different ways to improve perf. Limiting splat to the 255th or lower argument is a simple hack that gives good perf, and is good enough for an experiment. This PR series already has significant perf wins in rust-lang/rust#155223 - [0.3% perf improvement across 45 primary benchmarks](rust-lang/rust#155223 (comment)). We're spending a small amount of that perf for the new feature in this PR. ### Out of Scope for this PR - Change codegen to de-tuple caller and callee - Better diagnostics - Full support for splatted function pointer arguments

teor2345 · 2026-07-01T03:06:28Z

@rustbot label +F-splat

teor2345 and others added 11 commits June 22, 2026 15:50

Limit splat AST and FnDecl to 255 args

a8c1a22

And remove redundant const fn

Impl HIR FnSig for #[splat]

1ac6b6a

Impl HIR typeck for #[splat]

b9d2e39

Remove a leftover direct call to fn_ctxt.label_generic_mismatches

79c9709

Fix ICE in label_fn_like when splatted arg index is out of bounds

b4f8649

Impl MIR lowering side-tables for #[splat]

bd1ba47

Impl actual MIR lowering for #[splat]

e5e24aa

Impl TypeInfo for splat

ec6ae17

Add UI tests for #[splat] unhappy path, generics, fn colors

abca2d0

- const, async, and unsafe functions - const generics, complex types, and where clauses (including impl Tuple)

Add failing test for #[splat] with dyn AsRef<T> where T: Tuple

ebcd34a

Add splat UI tests for 255 splatted arg limit

1d46627